Trump-Appointed Cybersecurity Chief Uploaded Sensitive Government Files to ChatGPT

The acting head of the U.S. government's top cybersecurity agency uploaded sensitive internal documents to ChatGPT, according to a Politico report.

Citing multiple officials familiar with the matter, the report said that Cybersecurity and Infrastructure Security Agency acting director Madhu Gottumukkala — a Trump appointee — uploaded contracting materials marked "for official use only" to the AI chatbot. The activity reportedly triggered several automated alerts designed to prevent sensitive government files from being improperly removed or exposed from federal systems.

Gottumukkala was granted special authorisation to use ChatGPT earlier in his tenure, when access to the tool was restricted for other CISA staff. Officials within the Department of Homeland Security, which oversees CISA, subsequently reviewed whether the uploads posed any risk to national security or resulted in unauthorised data exposure.

Security experts have long warned that submitting internal or restricted government documents to public AI systems can be dangerous. Even when materials are unclassified, uploading them to widely accessible large language models can risk unintended retention or future disclosure, particularly if the data is incorporated into model training or surfaced through subsequent user interactions.

A spokesperson for CISA told Politico that Gottumukkala's use of ChatGPT was "short-term and limited," and emphasised that the agency was assessing the scope and impact of the incident.

Before taking on his current role, Gottumukkala served as South Dakota's chief information officer under then-governor Kristi Noem. After joining CISA, he reportedly failed a counterintelligence polygraph examination — a test Homeland Security later described as "unsanctioned." Following that episode, Gottumukkala moved to suspend six career officials from accessing classified systems, according to the report.

The incident adds to growing scrutiny over how senior government officials are using generative AI tools — particularly within agencies responsible for safeguarding critical infrastructure and national cybersecurity.